Lance has often said (and also in this) that if P=NP that would be great for the world: much more efficient ways to build things, science could be done better, etc, and that is much more important than that modern crypto would no longer work. We now have the technology to do private key really well--- like a thumb drive that has a billion bits for 1-time pads.
I agree that the world would be better off in some ways, I wonder how much damage would be done in the transition period from public to private key. Would the world recover enough to reap the benefits of P=NP?
First think of what YOU would do if you showed P=NP (and lets assume your algorithm is either reasonable or could be made reasonable with some time and effort).
The novel Factor Man is about what someone who has solved P=NP does. I won't tell you how it goes, but they deal with the issue intelligently. So if I solved P=NP then I would first re-read it, and think through if I would do that, or modify what is done, or what. Its a good start.
I reviewed the book in SIGACT News or you can read my review here
On a slightly diff note, here is the latest argument I've heard for why P=NP:
Planar 2-coloring is in P
Planar 4-coloring is in P
Planar 3-coloring should be in P.
This was said by a very good math/cs ugrad at UMCP. I do not know if he was kidding.